The brain has been a mystery for generations. The authors of the project have always been curious to understand how the human brain works, how it perceives images and other senses and how it functions so wonderfully. Now finally, we are in an age where we can understand the underlying functioning of the neural structures within the brain and make an attempt to piece together why humans respond to situations the way they do. Both of us are from a Computer Science Background and the science of the brain has awed us. This project gives us an opportunity to combine our area of study (Artificial Intelligence) to our area of interest (Neuroscience) and this is our main motivation behind taking up this project.
The experiment that originated the dataset was designed to predict human brain activity associated with the meanings of nouns. Nine right handed subjects (aged between 18 and 32) participated in the experiment. Each subject was shown one word per trial and were asked to think about a set of properties associated with that word. Each word was presented for 3 seconds followed by a 7 seconds rest period. 5 words belonging to 12 categories were presented to each subject over 6 experimental epochs. An epoch is a setting where all the 60 words were presented, without repetition. Each epoch had all the 60 words but in a different order.
The pre-processed and analysed data was obtained from the following site: http://www.cs.cmu.edu/afs/cs/project/theo-73/www/science2008/data.html
The following classification algorithms were used to predict the category of word:
- Naive Bayes
- KNN
- Decision Trees
- Random Forest